Skimmed milk powder can be produced by different methods and is classified by the American Dry Milk Institute (ADMI) in different categories according to non-denatured soluble protein content as “low-heat”, “medium-heat”, and “high-heat” powder.1
Two physico-chemical criteria can be applied to classify powders on the basis of their denaturation by heat, i.e. solubility value and soluble protein.
1 Powder obtained at low, medium or high temperature.
3.1.1 Solubility value
The method used (modified ADMI method) consists of dissolving in cold form a given quantity (5 g) of powder in 35 ml distilled water. After shaking followed by centrifugation for 10 minutes at 1 800 rpm, the supernatant is removed; the deposit is put back in suspension in 40 ml distilled water and is subjected to further centrifugation for 10 minutes. The insoluble portion is then oven-dried at 95°C for 15 hours. The percentage of insoluble material can be calculated from the weight of the dry residue and consequently the solubility value.
3.1.2 Insoluble protein value
The method here consists in measuring the nitrogen content of the non-soluble part of a reconstituted milk under standard conditions (2 g powder in 80 ml distilled water). After shaking and centrifugation for 15 minutes at 3 000 rpm, the supernatant is removed and the sediment put back in suspension in 80 ml water. After further centrifugation the residue of insoluble matter is mineralized and the nitrogen is titrated by the Kjeldahl method. The insoluble protein value is calculated by comparing the nitrogen content of the residue to the total nitrogen content of 2 g powder.
Anhydrous milk fat can also show differences in quality due to the raw material, manufacturing process and length of storage.
Two chemical values are commonly used to describe the quality of this fat, i.e. peroxide value and acid value.
3.2.1 Peroxide value
Measurement of the peroxide value has been used to evaluate the degree of oxidation (AOAC, 1975). The method consists in dissolving the fat in 30 ml of a chloroform-acetic acid mixture. A small quantity of potassium iodide is added. The peroxide releases the iodine which is then titrated by sodium thiosulfate in a starch paste. The results are expressed in milliequivalents (meq.) per 1 000 g fat.
3.2.2 Acid value
The degree of lipolysis of fat is expressed by means of the acid value (IDF, 1959; IDF, 1978b). The method consists in dissolving the fat in an ether-alcohol mixture, then neutralizing the acidity of the fat by 0.1 N caustic soda in phenolphtalene. The results are expressed in ml of 0.1 N soda per 100 g fat.
A study was undertaken at the Polytechnic Institute of Lorraine (France) to determine accurately the effects on the taste of milk recombined from milk powder and butterfat of varying quality, and to establish the maximum and minimum limits of reincorporation of recombined milk in local raw milk.
The raw materials used in this study came from batches which were freshly manufactured during 1980.
Three skimmed milk powders, each with a different level of denaturation, were used. Their characteristics are shown in Table 1.
| Powder | Solubility (%) | Non-soluble protein (%) |
|---|---|---|
| Slightly denatured | 99.340 | 0.457 |
| Moderately denatured | 99.016 | 0.622 |
| Highly denatured | 97.292 | 1.668 |
In order to have a range of AMF qualities it was necessary to modify the original fat by oxidation or by controlled lipolysis. The characteristics of this basic substrate are as follows: moisture content 0.2 percent (measured by Karl-Fisher method); peroxide value 40 mg per 1 000 g fat; acid value 1.22 ml per 100 g fat. A portion of this basic sustrate was subjected to oxidation while another was subjected to lipolysis under the following conditions.
3.3.1 Oxidation of AMF
The unmodified fat was dissolved at 40°C and kept at that temperature during the whole oxidation period. The latter process continued for about 12 hours under the action of ultraviolet rays from a lamp placed about 30 cm from the surface of the fat. Under these conditions the peroxide value rose to 7.5 meq. per 100 g fat. This highly oxidized fat was then used in a mixture adjusted with unmodified fat to obtain different levels of oxidation. It was immediately added to milk that had been reconstituted or stored at 20°C away from light.
3.3.2 Lipolysis of AMF
The unmodified and previously dissolved fat was enzyme-lipolized (0.5 – 8 g lipase per 100 g fat). After shaking, the fat was kept at 36°C for periods varying from 24 h to 7 days in order to obtain different degrees of lipolysis. The fat was then used immediately for recombining the milk.
The raw milk (non-fat dry matter content 8.5 percent and fat content 3.5 percent) used for the purpose was subjected to pasteurization at 78°C for 20 seconds.
The recombined milk and the mixtures were made as follows:
Skimmed milk powder was poured slowly and without shaking in distilled water, heated previously to 60°C, at a rate of 89.5 g powder to 876.5 g water. Shaking was continued for three minutes and the reconstituted milk was stored at + 5°C for 14 hours to stabilize the protein hydration level. In this way 966 g of reconstituted milk was obtained.
The milk was then heated to 60°C and the fat dissolved at the same temperature was incorporated at the rate of 34 g per 966 g reconstituted milk. The mixture was homogenized in a 2-stage homogenizer (APV Junior) at 60°C and at 185 kg/cm2 pressure at the first stage and 75 kg/cm2 at the second stage. A recombined milk was then obtained having 3.5 percent fat and 8.5 percent non-fat dry matter.
The milk recombined in this way was then mixed in different proportions with the pasteurized fresh milk. Given the equal dry matter and fat matter composition of these different products, the composition of the mixed milk did not have to be changed; the latter was submitted to the tasting panel within a maximum period of two hours.
The tasters were persons qualified and trained to detect variations of fat taste in the recombined milk and of powder taste in the reconstituted milk. Their selection was made by a three-way test and by a classification test (AFNOR, 1976).
The sensory analysis locations were equipped as specified by French standards (AFNOR, 1972).
Between 5 and 10 samples were submitted per session and the average time of each testing was 20 minutes for each of the analysts.
The methodology was that of assigning points on a scale increasing with the intensity of each of the three sought-after flavours. The scale is shown in Table 2, below.
| Points to be assigned | Extent of defect |
|---|---|
| 1 | Nil |
| 2 | Uncertain |
| 3 | Perceptible |
| 4 | Considerable |
| 5 | Marked |
| 6 | Very marked |
To apply these results it was arbitrarily agreed that the threshold for perception of an unpleasant taste for 50 percent of the tasters should be set at an average number of 3 points.
Only one constituent at a time was changed during the test so as to assess the effect of the quality of the milk powder and the anhydrous fat on the taste of the reconstituted milk.
The analysis of milk powder quality is given in Table 3 together with the proportions used in the mixture to reach the level at which the defect called “powder taste” appears.
| Powder quality | Solubility (%) | Non-soluble protein (%) | Extent of mixture at level 3 (%) |
|---|---|---|---|
| Slightly denatured | 99.340 | 0.457 | 59.0 |
| Moderately denatured | 99.016 | 0.622 | 53.5 |
| Denatured | 97.292 | 1.668 | 45.5 |
By setting 3 points as the threshold of unpleasant taste perception, it was found that with a high temperature powder mixing should not exceed 45 percent if adequate organoleptic quality is to be maintained. On the other hand, with powder prepared at a low temperature it seems that up to 59 percent of recombined milk can be added. Furthermore, Figure 1 shows the extent of the defect as a ratio of the percentage of recombined milk included in raw milk. According to the extent of powder denaturation, the panel described the taste defect of recombined milk not mixed with raw milk as marked or considerable.
Figure 1. Influence of rate of incorporation of recombined and raw milk on defect referred to as “powder taste”.
The influence of anhydrous milk fat quality is demonstrated by the appearance of an oxidized taste and a hydrolytic rancid taste.
The extent of the oxidized taste shows a linear correlation with the amount of recombined milk added (Figure 2). As the peroxide value rises, the defect increases.
Figure 2. Influence of fat oxidation and of rate of mixture of recombined and raw milk on the defect called “oxidized taste”.
There is a correlation between the fat peroxide value and the acceptable level of re-incorporation of recombined with raw milk as shown in Figure 3.
Figure 3. Maximum level of addition of recombined milk in raw milk (level 3 of sensory analysis) as a function of the fat peroxide value.
In the case of the original (unmodified) fat, it seems therefore that more than 50 percent recombined milk can be added. On the other hand, in the case of a fat with a very high peroxide value (7.5 meq) the maximum possible level of addition is very low (less than 10 percent).
These trials indicate that the measurement of peroxide value is useful in ascertaining the extent of the oxidized taste in milk. Statistical analysis of the findings has shown that the correlation is significant (r > 0.88).
Perception of a hydrolytic rancid taste depends on the incorporation rate of recombined milk with raw milk in a similar manner as perception of oxidized taste. In testing for the maximum and minimum degree of addition resulting in a detectable objectionable taste, a correlation may be seen between this rate and the acid value expressed in ml 0.1 N (caustic) soda. The correlation is linear within the limits of the experiment, in other words between acid values of 1.2 and 8 ml (Figure 4). The correlation coefficient is highly significant (r = 0.99) and warrants the use of acid value measurement to predict the degree of the rancidity defect. For example, with fresh fat (acid value 1.2) it seems that up to 70 percent of recombined milk can be introduced in raw milk without changing its organoleptic character. On the other hand, if the fat is very rancid (acid value 8) the maximum level is very low (less than 10 percent).
Figure 4. Maximum level of addition of recombined milk in fresh milk (level 3 of sensory analysis) according to fat acid value.
With the sensory method employed it was possible to identify the appearance levels of defective taste resulting from the use of raw materials with different degrees of spoilage. Thus use of a slightly denatured powder of the low temperature type warrants a higher rate of incorporation than with highly denatured powders. However, the differences remain small (45 to 59 percent under test conditions). It is pointless, therefore, even with a high quality powder, to try to produce a milk having a flavour equal to that of fresh milk.
Taste defects resulting from an alteration of fat have a much more marked influence on levels of mixing with fresh milk. It is necessary therefore to have complete control over the chemical quality of anydrous milk fat (its peroxide and acid values should be as low as possible). The generally accepted quality standards make it difficult to have recombined milk reincorporation percentage in excess of 40–50 percent.
It should also be pointed out that the group of tasters is not necessarily representative of the consumers from whom the recombined milk is intended. Because of food habits, differences in perception levels are likely so that the results given in this report should be interpreted and transposed with caution and the particular conditions of the respective population groups borne in mind. It would be useful to correlate these results with estimates based on surveys conducted among local populations.
